the word 'number' after 'the'
A neuronOne of thousands of tiny dials inside each layer. It reads the text so far and produces a number, its activation, that nudges what the model writes next. in layerOne step in the stack the text passes through. Qwen3-0.6B has 28 layers, each with 3,072 neurons. 22 of Qwen3-0.6B. Its strongest snippets are very consistent: they mostly peak on the same token. It is featuredA neuron picked for the full treatment: steering outputs and a written explanation. About 3,000 of the 86,000 neurons are featured., so it has steering outputs.
What this neuron does
This neuron responds most to the word 'number' right after 'the', as in 'the number of samples' or 'the number of correct answers'. Turning it up produced rambling text that did not show a clear link to that word.
Written by Claude from this neuron's activations and steering outputs. Confidence: medium. Think it is wrong? Flag it (opens a prefilled issue on GitHub).
Fires most on
The text windows where this neuron's activationThe number a neuron produces for a token. Large positive means it fired hard; negative means it pushed the other way. was highest. Amber marks positive activation; the outlined tokenA piece of text the model reads at once: often a word, sometimes part of a word or a punctuation mark. is the peak. Hover a token for its strength.
-
of 0.02). Plug your population size and required margin of error into the formula. The result will be the number of samples you need to take.⏎For example, suppose that you have a group
-
help judge the effectiveness of the course, Jessica measured the number of words per minute she could read prior to enrolling in the course. She obtained the following five number summary:⏎Use this information to draw a boxplot of the reading speed.⏎Answer
-
’s First⏎Before you learn how to calculate and monitor your target training heart rate, you have to know your resting heart rate. Your resting heart rate is the number of times your heart beats per minute while
-
, mathematical textbooks typically refer to nodes as vertices. They are also more precise in mathematical notation. For example, this chapter will use the symbol N to denote the number of nodes in the graph. In a mathematics context, the set of nodes is
-
number is to find the value of the mark. You can do this by using a percentage calculator. You can
-
f(), and ldexpl() return ±HUGE_VAL, ±HUGE_VALF, and ±HUGE_VALL (according to
SteeringAdding a push along a neuron while the model writes, to see how its output changes. Turning it up or down.
Push this neuron up or down while the model writes, and watch what changes. The outputs were generated ahead of time, so nothing runs in your browser.
Here is something I have been thinking about lately: the power word "power" — the power word power word power word power word power word power word — the word power word word power word word power word power word power word word power word word word power word word word word power word word
Most negative on
Where the activation went furthest below zero. Blue marks negative activation. Many neurons mean something different in each direction.
-
Information:||Web: http://www.tandfonline.com/pricing/journal/VECE20|⏎When requesting a correction, please mention this item's handle: RePEc:taf:jeduce:v:42:y
-
that the groups with the most weight loss were the diet-only and diet and exercise group with women in those groups losing an average of 10 percent of their starting weight, according to the study.⏎Additionally, the women
-
directly to that person rather than through a companion or interpreter. Conduct interviews in a manner that emphasizes abilities, achievements and individual qualities.⏎Address people who have disabilities by their first names only when calling everyone present by
-
-2 viruses were diverse and changed over time, but that the percentage of infected participants that was<|endoftext|>If you are participating with a teacher, please enter your teacher's code here:⏎An email has been sent to you (or your parent if you
Connections
Show on mapWhich neuronsOne of thousands of tiny dials inside each layer. It reads the text so far and produces a number, its activation, that nudges what the model writes next. in earlier layersOne step in the stack the text passes through. Qwen3-0.6B has 28 layers, each with 3,072 neurons. feed this one, and which later ones it feeds. Amber links excite, blue links suppress; longer bars are stronger.
Fed by earlier layers
- L19:603 suppresses
- L19:2761 suppresses
- L18:2718 excites
- L19:1176 excites
- L19:1928 excites
- L20:996 excites
- L18:2765 excites
- L19:870 excites
- L18:1304 suppresses
- L20:1284 suppresses
- L18:1555 excites
- L19:1378 suppresses
- L19:862 excites
- L20:706 suppresses
- L18:2049 excites
- L16:1676 excites
- L16:2258 suppresses
- L19:1514 excites
- L20:2314 suppresses
- L14:258 excites
- L17:1979 excites
- L19:1233 excites
- L21:2555 excites
- L21:449 suppresses
Feeds later layers
- L27:2084 suppresses
- L23:94 excites
- L26:2244 suppresses
- L27:612 weakly suppresses
- L24:78 weakly suppresses
- L27:425 weakly suppresses
- L25:417 weakly suppresses
- L24:574 weakly suppresses
- L25:1651 weakly suppresses
- L24:1374 weakly suppresses
- L23:926 weakly suppresses
- L23:134 weakly suppresses
- L27:1143 weakly suppresses
- L23:2778 weakly suppresses
- L25:1391 weakly suppresses
- L24:418 weakly suppresses
- L25:2922 weakly suppresses
- L23:817 weakly suppresses
- L27:1752 weakly suppresses
- L25:1114 weakly suppresses
- L24:121 weakly excites
- L26:507 weakly suppresses
- L27:393 weakly suppresses
- L25:210 weakly suppresses
Activation histogramA chart of how often a neuron takes each activation value.
How often this neuron fires at each strength. Almost always it does close to nothing (the tall bars in the middle); the snippets above are the rare moments it fired hard, amber for positive and blue for negative.
Logit lensReading which output tokens a neuron pushes toward or away from, straight off its output weights, before the later layers have their say.
The words this neuron nudges the model toward (amber) or away from (blue) when it fires. A hint, not a label: many neurons work through later layersOne step in the stack the text passes through. Qwen3-0.6B has 28 layers, each with 3,072 neurons. instead.
Promotes
- dating
- cmds
- otherwise
- of
- пут
- gratuites
- Cars
- OfString
- śmy
- abra
Suppresses
- ilar
- autical
- ycl
- 院
- ctal
- pirit
- :relative
- mlink
- sembled
- nces